12 PAPERS, ETC. 
than about 1,700 churches, though we know that many 
more, and some (as Dorchester, in Oxfordshire,) of great 
importance, existed in places where no church is men- 
tioned. 
A very few years ago we knew but little more of 
Norman work than we do now of Saxon, and in those few 
years we have learned to discriminate between early and 
late Norman, and to decide with tolerable accuracy the 
date of each building. Nor are differences wanting in 
Saxon work which may probably be discovered to be 
good criteria of date. For instance, may we not conclude 
from the decidedly Roman character of Brixworth, and 
the Roman composition of the cement at Britworth, that 
they are of very early date—either relics of Roman civili- 
zation, or instances of the skill directly imported from 
Italy. Again, if we find, as we very probably may, that 
in some cases the pilaster-like strips of stone are really 
superficial, instead of extending through the thickness of 
the wall and forming a frame-work for the rubble,—may we 
not, taking the analogy of Grecian architecture for our 
guide, conclude that those buildings in which only the 
appearance exists, are of later date than those in which 
the strips are actually the frame of the building; an idea 
which appears to me to gain strength from the fact of the 
long and short stones in the quoins of Sompting tower, 
(which, from its similarity to that described by Wolstan, is 
probably not earlier than the middle of the tenth century, 
being of equal breadth, and consequently not really useful 
as bonding ; and when we find plans or details of marked 
Byzantine character, may we not suppose that they were 
introduced either by the Northmen, (whose intimate con- 
nection with Constantinople is proved by the fact of 
Harold Hardrada commanding the Varangian Guard 
about the year 1030, as well as by the coins of Greek 
